Methylation of histone H3 at Lys-9 is an early mark on the X chromosome during X inactivation.

Coating of the X chromosome by Xist RNA is an essential trigger for X inactivation. However, little is known about the early chromatin remodeling events that transform this signal into transcriptional silencing. Here we report that methylation of histone H3 lysine 9 on the inactive X chromosome occurs immediately after Xist RNA coating and before transcriptional inactivation of X-linked genes. X-chromosomal H3 Lys-9 methylation occurs during the same window of time as H3 Lys-9 hypoacetylation and H3 Lys-4 hypomethylation. Histone H3 modifications thus represent the earliest known chromatin changes during X inactivation. We also identify a unique "hotspot" of H3 Lys-9 methylation 5' to Xist, and we propose that this acts as a nucleation center for Xist RNA-dependent spread of inactivation along the X chromosome via H3 Lys-9 methylation.

A radiation hybrid transcript map of the mouse genome.

Expressed-sequence tag (EST) maps are an adjunct to sequence-based analytical methods of gene detection and localization for those species for which such data are available, and provide anchors for high-density homology and orthology mapping in species for which large-scale sequencing has yet to be done. Species for which radiation hybrid-based transcript maps have been established include human, rat, mouse, dog, cat and zebrafish. We have established a comprehensive first-generation-placement radiation hybrid map of the mouse consisting of 5,904 mapped markers (3,993 ESTs and 1,911 sequence-tagged sites (STSs)). The mapped ESTs, which often originate from small-EST clusters, are enriched for genes expressed during early mouse embryogenesis and are probably different from those localized in humans. We have confirmed by in situ hybridization that even singleton ESTs, which are usually not retained for mapping studies, may represent bona fide transcribed sequences. Our studies on mouse chromosomes 12 and 14 orthologous to human chromosome 14 show the power of our radiation hybrid map as a predictive tool for orthology mapping in humans.

Control of neurulation by the nucleosome assembly protein-1-like 2.

Neurulation is a complex process of histogenesis involving the precise temporal and spatial organization of gene expression. Genes influencing neurulation include proneural genes determining primary cell fate, neurogenic genes involved in lateral inhibition pathways and genes controlling the frequency of mitotic events. This is reflected in the aetiology and genetics of human and mouse neural tube defects, which are of both multifactorial and multigenic origin. The X-linked gene Nap1l2, specifically expressed in neurons, encodes a protein that is highly similar to the nucleosome assembly (NAP) and SET proteins. We inactivated Nap1l2 in mice by gene targeting, leading to embryonic lethality from mid-gestation onwards. Surviving mutant chimaeric embryos showed extensive surface ectoderm defects as well as the presence of open neural tubes and exposed brains similar to those observed in human spina bifida and anencephaly. These defects correlated with an overproduction of neuronal precursor cells. Protein expression studies showed that the Nap1l2 protein binds to condensing chromatin during S phase and in apoptotic cells, but remained cytoplasmic during G1 phase. Nap1l2 therefore likely represents a class of tissue-specific factors interacting with chromatin to regulate neuronal cell proliferation.

Physical mapping of the autoimmune disease susceptibility locus, Bphs: co-localization with a cluster of genes from the TNF receptor superfamily on mouse chromosome 6.

An important approach to understanding complex diseases is to reduce them into well-characterized subphenotypes that are under monogenic control. One such example is Bordetella pertussis toxin-induced histamine sensitization in mice, a subphenotype of experimental allergic encephalomyelitis and experimental allergic orchitis. This subphenotype is controlled by a single locus, Bphs, previously mapped to a 33 cM region on mouse Chromosome (Chr) 6. We achieved considerable reduction of this candidate region and constructed a YAC contig across the refined interval. Our results demonstrate that Bphs is located between D6Mit151 and a newly developed marker, EC108RR, a region containing a small cluster of genes belonging to the TNF receptor superfamily. Sequence and quantitative analysis of the candidate gene, tumor necrosis factor receptor 1 (Tnfr1, p55), indicates that it is unlikely to be Bphs. However, the location of Bphs, together with physiologic effects it shares with Tnfr1 activation, suggest that Bphs may prove to be another member of the TNF receptor superfamily.

Linkage disequilibrium and physical mapping of Pas1 in mice.

By using linkage disequilibrium (LD) analysis in 21 strains of known susceptibility to lung cancer and by assembling a YAC contig, we mapped to a approximately 1.5-Mb region on distal mouse chromosome 6 the Pas1 locus, the major determinant of lung cancer predisposition in mice. Our results, on the basis of haplotype and phenetic analysis, suggest that the Pas1(s) susceptibility allele is shared by several mouse-inbred strains of independent origin, which show either high or intermediate predisposition to lung tumorigenesis. Therefore, the Pas1(s) allele is probably derived from an ancestral mouse rather than from independent mutations of the same gene. We showed the feasibility of LD in common inbred strains for the fine mapping of disease loci, and provided the biological basis and the reagents for the cloning of the Pas1 gene.

A new family of mouse genes homologous to the human MAGE genes.

The human MAGE genes are expressed in a wide variety of tumors but not in normal cells, with the exception of the male germ cells, placenta, and, possibly, cells of the developing embryo. These genes encode tumor-specific antigens recognized by cytolytic T lymphocytes. The MAGE genes are located on the X chromosome, in three clusters denoted MAGE-A, B, and C, mapping at q28, p21.3, and q26, respectively. The function of these genes remains unknown. Because mice offer many advantages for the study of genes that may be involved in embryonic development, we looked for the murine equivalents of the 12 human MAGE-A genes. Using a MAGE-A probe, we isolated 8 new murine genes that are homologous to the MAGE genes. On average, the open reading frames (ORFs) of these 8 closely related genes display a slightly higher degree of nucleotide identity with the MAGE-A ORFs than with the MAGE-B or MAGE-C ORFs. Furthermore, like MAGE-A genes, they encode acidic proteins, whereas the MAGE-B genes encode basic proteins. Accordingly, these 8 murine genes were named Mage-a1 to 8 (approved symbols Magea1 to 8). Mage-a genes were mapped in two different loci on the mouse X chromosome. Mage-a4 and Mage-a7 are located in a region that is syntenic to either Xp21 or Xq28. The 6 other genes are arranged in a cluster located in a region syntenic to Xp22. Like their human counterparts, Mage-a genes were found to be transcribed in adult testis, but not in other tissues. Expression of some Mage-a genes was also detected in tumor cell lines. Two Mage-a genes were found to be expressed in blastocysts.

Role of the region 3' to Xist exon 6 in the counting process of X-chromosome inactivation.

During early embryogenesis of female mammals, one of the two X chromosomes is randomly chosen to be inactivated in each cell, leading to the transcriptional silencing of thousands of genes on this chromosome. This random X-inactivation process also occurs during in vitro differentiation of female embryonic stem (ES) cells. A locus on the X chromosome, the X inactivation centre (Xic) is initially 'counted', given that at least two copies of Xic must be present per diploid genome in order for inactivation to occur. The counting process ensures that one X chromosome remains active in diploid cells. In the mouse, the essential functions of Xic can be assured by a 450-kb region containing the Xist gene. Xist maps within Xic (refs 7-10) and is necessary in cis for inactivation. The Xist transcript is a 15-kb RNA which is confined within the nucleus and coats the inactive X chromosome. In order to characterize functional elements within Xic and the Xist gene, we created a 65-kb cre/loxP deletion extending 3' to Xist exon 6. In undifferentiated ES cells, Xist expression from the deleted X chromosome was markedly reduced. In differentiated XX ES cells containing one deleted X chromosome, the X inactivation process still occurred but was never initiated from the unmutated X chromosome. In differentiated ES cells that were essentially XO, the mutated Xic was capable of initiating X inactivation, even in the absence of another Xic. These results demonstrate a role for the region 3' to Xist exon 6 in the counting process and suggest that counting is mediated by a repressive mechanism which prevents inactivation of a single X chromosome in diploid cells.

The mottled mouse as a model for human Menkes disease: identification of mutations in the Atp7a gene.

Mutations in the Atp7a gene, the mouse homologue of the MNK (ATP7A) gene, have been suggested to be responsible for the mottled phenotype. To date, despite considerable effort, changes associated with the mottled mutations have been detected in only two such mutants. In this study, we identify changes in the level of Atp7a transcript and mutations which could explain the mottled phenotype in nine out of the 10 mutants analysed. The fluorescence-assisted mismatch analysis method used here has proved particularly well suited for mRNA scanning of heterozygous carrier animals, because of its ability to detect mutations even in the presence of an excess of wild-type mRNA. The three new underlying mutations identified at the Atp7a locus include a splice mutation and two missense mutations. While the spectrum of mutations detected in the Atp7a murine gene provides an explanation for at least part of the wide phenotypic variation observed in mottled mutant mice, there is a singular absence of deletions which are associated with a sizeable fraction of human Menkes syndrome cases.

Genomic organization of the mottled gene, the mouse homologue of the human Menkes disease gene.

The mouse homologue of the Menkes gene has been shown to span 120 kb of genomic DNA and to be similar in structure to both its human MNK homologue (ATP7A) and the Wilson disease gene (WD; ATP7B). Conservation of the majority of intron/exon boundaries among the three genes was also observed. The high overall conservation of both the Atp7a gene and the direction of transcription of the Atp7a, Pgk1, and Xnp genes between human and mouse is compatible with the evolution of an ancestral gene subject to strong evolutionary constraints lying within a locally relatively conserved region of the X chromosome.

Structure, chromosomal location, and expression pattern of three mouse genes homologous to the human MAGE genes.

The human MAGE1 gene directs the expression of an antigen recognized on a melanoma by autologous cytolytic T lymphocytes. MAGE1 belongs to a family of genes that are expressed in a number of tumors of various histological types but not in normal tissues except testis. The MAGE genes are arranged in two groups that are located within two different regions of the human X chromosome (Xq26-qter and Xp21.3). By hybridizing mouse genomic libraries with a MAGE1 probe, we identified three homologous genes. Two of these mouse genes, Smage1 and Smage2, are more than 99% identical to each other and encode the same protein of 330 aa. The 5' noncoding region of Smage2 provides the potential for regulating the expression of the gene through several different promoters located in front of alternative first exons. The third gene, Smage3, has the structure of a processed transcript. It codes for a protein with only 11 aa substitutions with respect to the Smage1/2 product. Somatic cell hybrids and interspecific backcross analysis showed that Smage3 is autosomal and that Smage1 and Smage2 are located between the Dmd and the Ar loci on the mouse X chromosome. Since this region is syntenic to the human Xp21.1-p22.1 region, we conclude that Smage1 and Smage2 are homologous to the MAGE-Xp rather than to the MAGE-Xq genes. Smage1/2 transcripts were detected in several tumor and embryonal cell lines but not in normal mouse tissues with the exception of testis. Expression of Smage3 was found in embryos from Day 11 to Day 15.

Structural analysis of mouse glycine receptor alpha subunit genes. Identification and chromosomal localization of a novel variant.

The inhibitory glycine receptor is a ligand-gated ion channel protein that occurs in different developmentally regulated isoforms in the mammalian central nervous system. Here, we have analyzed genomic clones covering the coding regions of the murine glycine receptor alpha 1 and alpha 2 subunit genes. Both genes contain eight intronic regions with precisely conserved boundaries. The same structure was also found for seven exons of a third homologous gene, alpha 4, identified during screening. The predicted alpha 4 polypeptide displays very high homology to the alpha 2 subunit. Like the alpha 2 gene, the alpha 4 gene maps to the mouse X chromosome. Our data indicate that the genomic organization of glycine receptor alpha subunit genes is conserved during evolution.

Physical mapping and YAC contig analysis of the region surrounding Xist on the mouse X chromosome.

The Xist sequence has been proposed as a potential candidate for the X-inactivation center based both on its localization within the candidate region for the X-inactivation center in man and mouse and on its unique pattern of expression from the inactive X chromosome. We have cloned 550 kb of DNA surrounding the mouse Xist sequence in contiguously overlapping YAC clones and have developed a long-range restriction map that spans almost 1 Mb of this region and includes this YAC contig. The detailed restriction map we have established provides a framework for the identification of expressed sequences other than Xist that may equally exhibit unusual expression characteristics associated with X inactivation. The presence of possible structural or methylation differences within this region between the active and inactive X chromosomes has been investigated through comparative analysis of male and female genomic DNA, and we report here the identification of certain CpG-containing restriction sites around Xist that have an interesting differential methylation status on the inactive and active X chromosomes.

The XLR sequence family: dispersion on the X and Y chromosomes of a large set of closely related sequences, most of which are pseudogenes.

The XLR sequence family encodes RNA transcripts specific to late-stage T and B cells and their neoplasms. Only one apparently functional mRNA has been identified thus far and this encodes a novel 25 kDa nuclear protein. In this report, we find that the XLR gene family is composed of 50-75 copies per haploid genome which localize to at least two different portions of the mouse X chromosome. Neither of these locations are near the xid mutation that earlier work had correlated with XLR. In addition, some members of this family are also on the Y chromosome. Another surprising finding is that while the fourteen genomic clones examined to date have the same exon-intron structure and are closely related with respect to sequence conservation (90%), all appear (in most cases by multiple criteria) to be non-functional, raising the possibility that all but one of the members of this large semi-dispersed family are pseudogenes.

A-raf oncogene localizes on mouse X chromosome to region some 10-17 centimorgans proximal to hypoxanthine phosphoribosyltransferase gene.

The localization of the A-raf cellular oncogene on the mouse X chromosome has been determined using Xbal-restricted DNAs prepared from progeny of an interspecies backcross between the B6.CBA.R1 and the Spe/Pas mouse strains. This localization to the proximal part of the mouse X chromosome has been confirmed by the use of somatic cell hybrids, carrying partially deleted X chromosomes and suggests that the A-raf oncogene localizes to a region lying some 10-17 centimorgans proximal to the hypoxanthine phosphoribosyltransferase (Hprt) gene between the locus DXPas4 and the locus DXPas7 defined by the cross-reacting human X chromosome-specific probe DXS32 (M2C). This localization on the mouse X chromosome is compatible with the presence of the A-raf oncogene on the short arm of the human X chromosome between the centromere and Xp21.

Minor histocompatibility antigens are developmentally regulated on murine embryonal carcinoma cells and their early differentiated derivatives.

Differences in the expression of minor histocompatibility (Hm) alloantigens on two mouse embryonal carcinoma (EC) cell lines and the PYS-2 and T.D.M.-1 differentiated derivatives have been demonstrated by their ability to elicit a cytolytic T lymphocyte response. Experiments involving the use of various responder-target strain combinations and recombinant inbred mice strains have shown that: (1) there are major differences in Hm expression on EC cells compared with differentiated derivatives whose Hm expression appears more like that of adult splenocytes; (2) although both EC cell lines show reduced Hm immunogenicity compared with adult splenocytes, major differences in the expression and possible presentation of Hm between the F9 and PCC3 EC cell lines can be detected by in vivo priming and by in vitro cold competition target experiments. These observations are discussed in relation to the differences in allograft rejection patterns observed with PCC3 and F9 and to possible differences in developmental staging of these cell lines.

Developmentally regulated cell surface structures on mouse and human embryonal carcinoma cell lines.

Three monoclonal antibodies 5.1.H, 8.7.D and 13.7.A raised against semi-purified Tera 1 membrane fractions recognize distinct onco-foetal antigens which are developmentally regulated on cells such as Tera 2 clone 13 and appear to be restricted in their expression to undifferentiated ectoblastic cells and certain organized cystic structures mimicking the foetal intestine. These antigens, absent from normal adult tissues, differ markedly from glycosidic stage-specific antigens such as 75.12 which, while functioning as embryonal carcinoma differentiation markers, are also expressed on certain adult tissues. No evidence for a role of fucosyltransferases in regulating either 75.12 or SSEA-1 antigen expression on embryonal carcinoma cells or for the presence of lectin-like structures recognizing these antigens on such cells was found.

Differential Hm antigen expression on EC cells and early differentiated derivatives.

Differences in the expression of minor histocompatibility (Hm) alloantigens on two mouse embryonal carcinoma (EC) cell lines and the PYS-2 and T.D.M.-1 differentiated derivatives have been demonstrated by their ability to elicit a cytolytic T-lymphocyte (CTL) response. Experiments involving the use of various responder-target strain combinations on the one hand and Recombinant Inbred (RI) mice strains on the other have shown that: (i) there are major differences in Hm expression on the EC cells compared with the differentiated derivatives whose Hm expression appears more akin to that of adult splenocytes; (ii) although both EC cell lines show reduced Hm immunogenicity compared with adult splenocytes, major differences in the expression and possibly presentation between the F9 and PCC3 EC cell lines can be detected both by in vivo priming and by in vitro cold competition target experiments. These results are discussed in connection with the unexpected finding that some EC cell lines are capable of specific competition effects for appropriate CTL effectors despite their inability to stimulate such effectors in vitro and the absence of major histocompatibility complex (MHC) products.

Definition of three species-specific monoclonal antibodies recognizing antigenic structures present on human embryonal carcinoma cells which undergo modulation during in vitro differentiation.

Three novel species-specific monoclonal antibodies 5.1.H, 8.7.D and 13.7.A raised against semi-purified detergent solubilized fractions of the Tera 1 embryonal carcinoma (EC) cell line are restricted in their in vitro distribution to undifferentiated human EC cell lines. Competition experiments have established that distinct antigenic specificities are seen by the 3 different monoclonal antibodies. All 3 antigens 5.1.H, 8.7.D and 13.7.A, defined by these monoclonal antibodies, undergo developmental regulation and cease to be expressed on Tera 2 clones 5 and 12 after retinoic-acid-induced differentiation and on LICR LON HT 39/7 cells after phorbol-ester-induced differentiation. These results taken together with the extremely limited in vivo tissue distribution of the defined antigens suggest that the 5.1.H, 8.7.D and 13.7.A monoclonal antibodies define distinct onco-foetal antigens.